REM Sleep Classification with Respiration Rates

Abstract

Polysomnagraphy, the method evaluating the sleep quality, is the reliable and standard method. However, to perform it, subject has to attach many sensors to their body. It seems intrusive, tangled and interrupting the subjects sleep. For the sake of monitoring the daily sleep pattern, simple method measuring the biosignal nonintrusively is needed. At that point, respiration is a beneficial signal, because it can be easily acquired by check the change of the thorax or abdomen. Respiration is one of the representatives of the reflection of the autonomic nervous system. According to the transition of the sleep stages, autonomic nervous system also changes and it can be shown by respiration. In this study, REM and non-REM sleep is classified by only using respiration for the 3 subjects. To enhance the performance, activity signal acquired by load-cell installed bed is added to divide the wake and REM sleep. At the results, hypnograms with polysomnagraphic data is compared with them. It can be seen that high correlation between the REM sleep and increasing pattern of the respiration rates.